Project Summary Although military personnel may have vulnerabilities and susceptibilities for anxiety disorders similar to the non-military population, the extreme stressors of deployment and war time service enhance the likelihood of developing anxiety disorders such as post-traumatic stress disorder (PTSD). However, only a subset of individuals exposed to such stressors develops PTSD, leading to the concept of pre-existing vulnerability factors such that individuals with one or more of these vulnerability factors, who are then exposed to extreme stressors, are more likely to develop PTSD than individuals who receive equivalent exposure but have no (or fewer) vulnerability factors. Avoidance is a core symptom of PTSD; avoidance symptoms tend to increase rather than decrease over the course of the disorder, and may be particularly predictive of which trauma-exposed individuals are likely to develop PTSD. Vulnerability to PTSD may therefore partially reflect individual differences in an individual's pre-existing tendency to acquire, express, or maintain avoidant behaviors. Understanding the mechanisms by which such vulnerabilities translate into psychopathology in veterans is key for a better understanding of PTSD and also for designing effective therapeutic interventions that target these mechanisms. There is a rich tradition of studying avoidance learning in animals, including work from our lab and others, but work directly linking animal data to human avoidance has lagged behind. Here, we will employ a parallel program of computational modeling and human avoidance studies to begin to bridge the gap between human and animal studies. First, we will develop a computational model of avoidance learning to address data from animals, including data from a rat model of vulnerability to PTSD and anxiety that shows facilitated learning and persistence of avoidance learning. Second, in parallel with the modeling work, we will conduct studies of avoidance learning in veterans with varying degrees of PTSD avoidance symptoms, to determine whether veterans with severe avoidance symptoms show avoidance behavior in laboratory tasks that parallels the behaviors observed in the rats. These studies will provide valuable empirical data on avoidance learning in veterans, in addition to directly testing how closely the rat behavior mimics that observed in the humans. Third, we will apply the computational model to the veteran avoidance learning data, to determine whether the same computational mechanisms that sufficed to account for the rat data can also account for the human data. If so, this will increase our confidence in the translational potential of the rat model; if not, this will identify important limitations that can be studied in future animal and human work. In either case, the results from the computational modeling will begin to bridge the gap in our understanding of the mechanisms driving avoidance learning in rats and in veterans with and without PTSD avoidance symptoms. The long term goal of this program of work is to use computational modeling to better understand the processes by which vulnerability to avoidance translates into psychopathology in anxiety disorders including PTSD. Knowledge of these processes and their underlying mechanisms may guide future development of targeted therapies that modulate these mechanisms to affect the development and maintenance of aberrant avoidance, providing strategies to treat or prevent development of avoidance symptoms in PTSD.